CS Speaker Series

In Collaboration with LACOE's STEM Science Unit, USC Viterbi School Engineering  is happy to annnounce the 2022-2023 Computer Science Speaker Series!

These are monthly interactive, computer science webinars deisgned for K-12 teachers. You will hear from practicing computer scientists as they discuss their current research! In addition, they want to work with you to design a classroom experience for your students!

In follow up opportunities, you will be able to work with USC Professors and Grad Students to help you integrate Computer Science Content, as well as help develop activities that teachers and students can experience!

You are free to attend all or some of these webinars as your schedule allows!

September: Professor Robin Jia

Demystifying Deep Learning Methods for Text and Image Generation

Recent developments in deep learning, a subfield of machine learning that uses large-scale "neural network" models, have greatly improved the capabilities of computers to automatically generate text. These advances underpin notable natural language processing applications such as machine translation and chatbots. I will describe the basic machine learning principles behind how these models work, and show how text generation can be viewed as a sequence of classification problems. I will also demonstrate how the newest systems can "learn" to do new tasks and mimic "reasoning" to answer questions. Finally, I will discuss some possible concerns of these systems and the potential for harmful applications.

November: Professor Jesse Thomason

Natural Language Processing and Robotics

Computer agents that respond to natural language must be able to understand what a person wants them to do, such as Alexa playing a certain requested song, and even reason about the world around them, such as a vacuum robot told to clean the kitchen. We will discuss topics in computational language grounding: tying human language to information in the world and possible actions a computer agent could take in response.

February: Professor Vatsal Sharan

Fairness in Artificial Intelligence & Machine Learning Systems
Target Audience: High School

Artificial intelligence & machine learning (AI & ML) are increasingly being used in applications which involve significant societal implications and require high levels of trust in the system (such as in self-driving cars, or in deciding whether someone gets a loan). We will explore how to make AI & ML systems fair (such as not being biased against any subpopulations or demographics of individuals) and robust (such as still making good predictions when encountering different kinds of data), with an eye towards the societal applications.

October: Professor Somil Bansal

Safety in Robotics

From self-driving vehicles to autonomous drones, machine learning and AI are enabling robots to make intelligent decisions in novel situations and environments based on their past experience. However, they have also opened up a set of new challenges for robotics. Learning failures can lead to catastrophic robot failures and compromise human safety, as exemplified by recent self-driving car accidents. In this webinar, we will explore how we can empower our robots to understand their own limitations and make safer decisions as they work alongside humans and other robots.

January: Professor Feifei Qian

Robot Locomotion

Animals -- lizards, snakes, insects -- exhibit novel strategies in effectively interacting with their physical environments and generating desired responses for locomotion. In this webinar, I will discuss how my group's research integrates biology, physics, and robotics, to create robots that can do the same. Specifically, I will talk about how we extract general principles governing the interactions between locomotors (animals or robots) and challenging terrains, such as soft sand, deformable soil, and earthquake rubble. I will then discuss how we use these principles to create novel locomotion control strategies for robots to generate desired motion in these complex terrains.